Usage of Embedded Systems for DoS Attack Protection

Transcription

1 Usage of Embedded Systems for DoS Attack Protection I. Dodig 1, D. Cafuta 1, V. Sruk 2 1 Politechnic of Zagreb HR ZAGREB, I. Lucica 5, CROATIA {davor.cafuta, ivica.dodig}@tvz.hr 2 Faculty of Electrical Engineering and Computing HR ZAGREB, Unska 3, CROATIA vlado.sruk@fer.hr ABSTRACT An embedded system commonly represents special-purpose device usually based on processor designed to perform one or a few dedicated functions. Such systems are designed to maintain control over specific functions by gathering and reporting data received from sensors or some other devices. Embedded systems show increasing trend to modify data or control information flow, in some cases over the local area network or even Internet. Widespread network usage results in rapid development of advanced network connectivity in embedded systems. In this paper denial of service as a challenging problem in network security is introduced. Different defense methodologies are described. Further more, possibility of an embedded system advanced network connectivity usage in hardening network security is explored. Key words: Embedded system, Network, DoS I. INTRODUCTION Computer systems are today widespread over multiple aspects of human life. In some aspects of application, computer system can become invisible to surrounding. Fractions of those systems are devices designed to master some specific task called embedded systems. Small sized, low power consumption and economically cost-effective are the common features of embedded systems. Most commonly embedded system is realized as a single processor board without the operating system with single application performing given set of tasks. Application is limited to perform tasks conceived during the development process. Embedded system hardware can be specially designed or a generic solution can be used. Using generic hardware speeds up the development process and simplifies system upgrading but limits accessible hardware features and response times of the whole system. Development methodology needs to ensure cost-effective approach in design. Opposite to desktop computing embedded systems has better availability and security. An example of increasing availability in hardware design would be replacement of unreliable mechanical parts like hard disks and switches with electronic switches and solid-state disks. Another approach in embedded system design is a distributed system of embedded system units. Unit represents fully functional part of the system that collaborates with other devices to ensure the collective goal of the embedded system. This approach increases reliability introducing redundancy. In case of unit failure, other units can step in order to ensure functioning of the system. This approach requires implementation of failure detection and self-recovery in order to improve availability of the whole system. In this paper prospect of network attack detection using an embedded system is presented. Denial of service (DoS) attacks as security threats and some common defense methodologies are presented and network design using the embedded system is proposed. Second section presents common threats in network security. In third section, DoS attacks and common defense methodologies are presented. To increase network security different network designs strategies are described. Finally, defense solution is proposed by adding embedded systems in network design. II. NETWORK SECURITY ISSUES Internet has become the largest network of public data and information simplifying personal and business communication. Rapid Internet expansion induces rapid amplification in number of networks. Increasing data and information flow exponentially increases number of networks, Security threats, vulnerabilities and harmful attacks become common part in these networks. Hackers, viruses, revenges employees and human mistakes are some examples of the security risks. Security threats can become significantly serious if critical data can be changed or even deleted. Attacks can also harm network infrastructure or business profit disabling part or even the whole network. Increasing security measures often increases complexity of the system and decreases reliability of the system. Complexity of these measures can also effect on effectiveness or morale of employees, partners or clients [15,16,17]. This is especially important for business users, which are encouraged to create security policy to protect their businesses and clients from threats. Security policy is a document describing what it means to be secure for a system, organization or other entity. It incorporates company plans to protect the company's physical and information technology assets incorporating set of laws, rules, and practices for organization management, protection and sensitive information distribution.

2 There are three most common sources of attacks. Hackers used attacks to gain some advantage or to warn system administrator for problems in network security. Revenge employees presents second most common source. Their goal is to harm their firm by exposing data and preventing business tasks. Third common source represents users that inadvertently leave security vulnerabilities and enable other two sources to invade the network. Most common types of attacks use viruses or illegally installed remote administration programs known as Trojan horse. Remote administration software is a group of applications used by computer administrator to simplify computer configurations and administration over the remote networks. Trojan horse is often used to open a path for installing new viruses. Firewall protected secure computer networks can be affected with a virus only if it is spread from inside the network. Most commonly, user executes its code as it receives it by or downloads it from Internet. In such cases antivirus software should prevent virus code execution. Inadequately installed or lousy updated antivirus software significantly increases risk in contaminating other computers in closed networks. Network security attacks can be classified in three basic categories depending of the purpose of the attack. Reconnaissance attack is known as the attacks that gather data about network. Access attack is known as attack that tries to gain access to the network service by exploit the vulnerabilities. Prevent access to services or the whole computer system is the goal of Denial of service attack [16]. Reconnaissance attacks are usually first to be used by attacker to gain data used in other attacks. Applications like IP and port scanners and packet sniffers gather information and data about victim services. Using this information, data vulnerabilities can be searched and tools to exploit them can be programmed or downloaded from the Internet. To decrease and simplify development of the application and to accelerate application process programmers often neglect network security issues. Packet sniffers can catch packets from this kind of application containing unsecured data for example, usernames and password in plain text. Third person gathering these data can access various resources that can lead to complete computer system access. Denial of Service (DoS) attack is designed to prevent access to services or system overloading it with requests. These attacks produce large amounts of requests to devices that are part of the victim computer network thus blocking access to valid users: partners, employees or clients. The worst case of this attack is Distributed DoS attack. In these attack attackers, uses already compromised network devices (zombies) to increase the amount of the attack to the victim. In the next chapter DoS attacks are examined closely and some defense methodology will be examined. III. DOS ATTACK Most commonly used terms around network traffic are format and flow. Flow is describing traveling packets that carry data in IP networks. IP network are group of networks using the Internet protocol (IP) version 4 or version 6 depending on the form of the logical address. Format describes structure of the data, segments, packets, frames and bits. At the highest application level, data is used or generated by application. In the lower transport layer, data is segmented and marked to be recognized for which application is destined. Segments are encapsulated with logical source and destination address and are now known as packets on network layer. Packet is commonly referred to as a unit in networks. It is encapsulated in the frame with physical source and destination network at lower data link layer. Frame is changed whenever it passes the third layer network device (router). Finally, the frame is coded in bits to be processed by network interface card. At the transport layer when forming a segment reliable or unreliable delivery has to be used. Protocol used for reliable delivery over IP network is Transport control protocol (TCP). Opposite of TCP is UDP (user datagram protocol) providing unreliable delivery [17]. Nowadays, there is an issue of blocking unwanted packets on the Internet. One of the mayor issues are DoS attack packets which generate large amounts of network flow making networks unusable to end user. United States Computer Emergency Readiness Team (US CERT) has identified DoS attacks symptoms as: significant decrease in network performance making web sites and other network services become unavailable, significant increase in unwanted messages (spam). The last attack is known as mail-bomb [1]. An attack does not have to exclusively be formed as DoS attack. Attacks can use DoS as a part of the elaborate plan to overload the network to expose other vulnerabilities. DoS attacks can focus security issues to them enabling unnoticed malware installation. A DoS attack is characterized by an explicit attempt by attackers to prevent legitimate users of a service from using that service. Attacks can be directed at any network device, including attacks on routing devices and web, electronic mail, or Domain Name System servers. A DoS attack can be perpetrated in a number of ways. The five basic types of attack are: 1. Consumption of computational resources, such as bandwidth, disk space, or processor time 2. Disruption of configuration information, such as routing information. 3. Disruption of state information, such as unsolicited resetting of TCP sessions. 4. Disruption of physical network components

3 5. Obstructing the communication media between the intended users and the victim so that they can no longer communicate adequately. A DoS attack may include execution of malware intended to max out the processor's usage, preventing any work from occurring, trigger errors in the microcode of the machine, trigger errors in the sequencing of instructions to force the computer into an unstable state or lock-up. Basic type of DoS attack is presented in figure 1. Attacker from the Internet is creating large amount of unwanted packets directly to victim Adding intermediators and agents between attacker and victim transfer DoS attack to DDoS. DDoS attack is presented in figure 2. Attacker is using one or more intermediate to send request to agents. Finally, an agent sends request to the victim. Agents and intermediator are most commonly computers infected with viruses or Trojan horses. These computers in literature are known as Zombies and are infected with computer programs designed for the purpose of generating attacks. This tree like structure significantly increases overload and consequences on the victim computer system. DDoS significantly complicates finding the source of the attack and thus complicate the defense strategies [3]. Uninfected computers can also be used as attackers. Agents send requests with faked logical source address of the victim and destination address of the uninfected computer known as reflector. Reflectors responses accoring to the IP protocol to faked logical address which results in overloading the victim with requests. This type of attack is known as Distributed Reflective Denial of Service (DRDoS) [4]. Attack can be also carried out exploiting specific vulnerability in application or network protocol. Application vulnerabilities are most common types of attack that can be eliminated as the applications are often updated or replaced. Cisco 7xx routers have exploits in some version of operating system which manifest in complete shutdown in case of trying to connect through telnet using very long password. Other examples of attacks using the application based vulnerabilities of operating systems are Ping of Death[5], Land[6], TearDrop[7]. Examples of attack based on application software are Finger bomb[8] and Chargen attack[9]. Exploiting network protocol weakness lead to DNS cache poisoning[10][11] attack which files DNS cache memory with incorrect data. Other groups of DoS attack are based on starving the victim with valid client requests. An example of these attacks would be Ping Flooding[7], SYN Flooding[12] and UDP Flooding[13]. Most commonly referenced starving DoS attack is SYN Flooding attack. This attack exploits the vulnerability of the TCP network protocol. This type of attack constitutes with 50% in DoS starving attacks. Figure 1. An exaple of basic DoS atatck A characteristic example of vulnerability in TCP network protocol is a three-way handshake. It is used to form a TCP connection between two parties: 1. Client is sending a SYN message to the server 2. When server receives the SYN message it reserves requested resources and sends SYN- ACK message back to client. 3. Client confirms receive of SYN-ACK by sending an ACK message. Upon receiving the ACK message server establishes the connection. If in last step attacker never send an ACK message, server has to hold resources for certain time before freeing it to the system. As the amount of system resources is known to be limited after multiple open request all server resources can be blocked and thus server would have to refuse any requests for connection for significant amount of time. In case of DoS attack, the investigative process should begin immediately after the DoS attack begins. There will be multiple phone calls, callbacks, s, pages and faxes between the victim organization, one's provider, and others involved. This can be a very time consuming process. It has taken some very large networks with plenty of resources several hours to halt a DoS attack. The easiest way to survive an attack is to have planned for the attack. Having a separate emergency block of IP addresses for critical servers with a separate route can be invaluable. A separate route (perhaps a DSL) is not that extravagant, and it can be used for load balancing or sharing under normal circumstances and switched to the emergency mode in the event of an attack [14]. Filtering is often ineffective, as the route to the filter will normally be swamped so only a trickle of traffic will survive. However, by using an extremely resilient statefull packet filter that will inexpensively drop any unwanted packets, surviving a DoS attack becomes much easier. When such a high performance packet filtering server is attached to an ultra-high bandwidth connection (preferably an Internet backbone), communication with the outside world will be unimpaired so long as not all of the available bandwidth is saturated, and performance behind the packet filter will remain normal as long as the packet filter drops all DoS packets. It should be noted, that in this case the victim of the DoS attack still would need to pay for the excessive bandwidth. Firewalls have simple rules such as to allow or deny protocols, ports or IP addresses. Some DoS attacks are too complex for today's firewalls, e.g. if there is an attack on port 80, firewalls cannot prevent

4 that attack because they cannot distinguish good traffic from DoS attack traffic. Additionally, firewalls are too deep in the network hierarchy. Routers may be affected even before the firewall gets the traffic. Nonetheless, firewalls can effectively prevent users from launching simple flooding type attacks from machines behind the firewall. Some firewall present built-in capability to differentiate good traffic from DoS attack traffic. This capability is known as a "Defender", as it confirms TCP connections are valid before proxying TCP packets to service networks (including border routers). Other measures include limiting number of incoming connections from IP range, protocol analysis, checksum verification (so that the packets are not altered since transmission). Routers and switches also include options for managing connection speed and ACL (Access Control List) as firewall. Servers are usually equipped with Intrusionprevention systems (IPS) but they are effective if the attacks have signatures associated with them. However, the trend among the attacks is to have legitimate content but bad intent. These defense methodologies can not prevent DoS attacks by themselves. Combining several methods can gave better results. In this paper proposal in network design is made to integrate embedded system as defense line from DoS attacks. IV. ATTACK PREVENTION USING EMBEDDED SYSTEM Computer network is a group of directly connected computers. Advantages of using computer networks include information and resource sharing in large scale. First computer network was designed by the Ministry of Defense of the United States of America in 60 s. This network is known under name ARPANET (Advanced Research Projects Agency Network) designed by ARPA (Advance Research Projects Agency). Computer network has evolved and become available all over the world. To adapt to hardware challenges, different topologies and protocols were designed and adopted. Bus, ring and star topology are the most known network topologies. As the network grows, security becomes issues. Simple network design is presented in figure 3. Presented architecture does not have any protection from attacks. If attack is executed on this architecture, network and their resources will become inaccessible to other users. Network consists of the router directly connected to Internet over the ISP (Internet service provider). If the router lacks any DoS attack protection server would be attacked and very likely become unreachable. In some cases, the server of the victim does not have to be goal of the attack. The the goal can be a victim router to become zombie in elaborate attack on another victim. Network security upgrade can be achieved by introducing firewall between router and server. Firewall has to be equipped with several DoS defense strategies. Architecture is presented in figure 4. Nowadays, this is the most common network architecture. Disadvantage of this system is lack of any protection of the router. Router can be attacked from the Internet and blocked. Amount of network traffic can even damage the router. Firewall protects the server and his resource but router blocks the valid network traffic to reach firewall and thus the server. Upgraded network design with an embedded system is presented in figure 5. Task of the embedded system is to detect attempt of DoS attack by analyzing and resolving packages. To be able to carry this tasks system is equipped with database support. This architecture is very similar to those in figure 4. All of figure 4 advantages are successive. Main drawback of figure 4 is solved. In case of DoS attacks it would be stopped in front of ISP router saving router from unwanted traffic. Embedded system can be at the back of the router in case that ISP does not allow changes in network architecture in front of the router. Advantage of using an embedded system is by executing a precise task of defending against DoS attacks. Firewall behind the embedded system does not have to worry about this attack and can perform his tasks more efficiently. Due to simple hardware and no other process to occupy embedded system should perform better than regular server with installed application against DoS attacks. Reliability of the system can be significantly increased by using the embedded system. Some issues that have to be addressed in design are overloading in case of DDoS or DRDoS attacks and algorithms for managing database information and network packets flow. Figure 3. Unprotected network architecture Figure 4. Network architecure with Firewall

5 Embodying an embedded system device in network architecture, other network devices like firewall and routers benefits in performance and functionality. This way firewall and routers may well focus to their primary tasks. Firewall itself should not be the first line of defense in the case of DoS attacks. In such cases decreasing time of threat detection and preventing the attack sun as possible is of crucial importance to achieve both adequate network performance and security. Figure 5. Network architecure with Embedded system Methodoly of building such embedded system would be: Definition of possible DoS, DDoS, DRDoS attack Definition of maximal amount of traffic that system has to be designed to work with Hardware selection: generic system or specific hardware Software selection: using operating system or not Hardware and software verification System evaluation Hardware and software design Integration Verification These steps do not have to be serially executed. In attempt to create full proof embedded system some steps can be skippeed or repeated several times. V. CONCLUSION In this paper network architecture modification is proposed in order to increase protection of network parts against DoS related attacks by introducing an embedded system. DoS attacks are well-known and widespread security issues in computer networks. Methods and vulnerabilities used to perform these attacks are well known and documented. In addition, many solutions offer protection depending on the type of the attacks. Combining different solution on one device can become delicate, as this device should have sufficient resources to accomplish it. Our proposal to improve DoS attack prevention is introduction of the embedded system to simplify hardware and liberate it from other tasks requested by operating system or other system software of a desktop computer or server. An embedded system could be specialized hardware and software optimized thus achieving better performance than regular server or generic embedded system solution. REFERENCES [1] Mindi McDowell, National Cyber Alert System,, [2] Department of Homeland Security: Cyber Security Procurement, Language for Control Systems, [3] Rocky K.C. Chang, Defending against flooding based distributed denial of service attacks : A tutorial, IEEE Communications Magazine, October 2002 [4] Vern Paxson, An Analysis of Using Reflectors for Distributed Denial-of-Service Attacks, 2001, [5] CERT Advisory CA ,Denial-of-Service Attack via ping, html [6] Cisco Security Advisory: TCP Loopback DoS Attack (land.c) and Cisco Devices, December [7] Wong Natepetcharachai and Bo Zhang, DDoS Attack Tools and Incidents ~bozhang/ personal/pdfs/ddos.pdf [8] Internet Security System security_center/advice/exploits/services/finger/finger _bomb/default.htm [9] CERT Advisory CA : UDP Port Denial-of- Service Attack, html [10] Tom Olzak, DNS Cache Poisoning: Definition and Prevention, 2006, Papers/ DNS_Cache_Poisoning.pdf [11] DNS Cache Poisoning - The Next Generation, 2007, dnscache-poisoning/ [12] CERT Advisory CA TCP SYN Flooding and IP Spoofing Attacks, advisories/ca html [13] Indiana Ave, Advanced Networking Management Lab (ANML), Distributed Denial of Service Attacks(DDoS) Resources, 2008: [14] Corsaire White Papers, Surviving Distributed Denial of Service (DDoS) Attacks, 2004, p,00.htm [15] Computer and Network Security, [16] Network Security, warp/ public/cc/so/neso/sqso/ [17] Marin, G.A., Network Security Basics, Security & Privacy, IEEE, Nov.-Dec. 2005, Volume: 3, Issue: 6, On page(s): 68-72